- Email: [email protected]
Excimer Laser Ablation of the Cornea and Lens
Excimer Laser Ablation of the Cornea and Lens Experimental Studies CARMEN A. PULIAFITO, MD, ROGER F. STEINERT, MD, THOMAS F. DEUTSCH, PhD, FRANZ HILLENKAMP, PhD, ELLEN J. DEHM, BA, CATHERINE M. ADLER, BS, HT (ASCP)
Abstract: The pulsed ultraviolet excimer laser has been used to produce tissue ablation with a high degree of precision and with minimal thermal damage to adjacent structures. In comparative studies of excimer laser ablation of the cornea and crystalline lens using 193 nm and 248 nm radiation, threshold fluence for corneal and lens ablation was higher at 248 nm than at 193 nm. Ablation of corneal stroma at 193 nm produced the most precise cuts. When examined by transmission electron microscopy, a narrow zone of damaged tissue (0.1 to 0.3 Jlm) was seen immediately adjacent to the tissue removed by the laser. Ablation with 248 nm radiation produced incisions with ragged edges and with a wider and more severe zone of damage in adjacent stroma. Ultraviolet spectral transmission studies of the corneal stroma showed that absorption is 10 times greater at 193 nm than at 248 nm. The excimer laser was effective in producing well controlled ablation of the crystalline lens in vitro, with effects parallel to those seen in the cornea. [Key words: cornea, excimer, keratorefractive surgery, laser, lens.] Ophthalmology 92:741-748, 1985
Excimer lasers are high-power sources of pulsed ultraviolet radiation. Recently, 193 nm radiation from an argon-fluoride (ArF) excimer laser has been shown to produce precise etching of polymer films and clean incisions in biologic materials, in particular the cornea, skin and aorta. I - 6 The excimer laser has been proposed as a new laser therapeutic device with the potential of
From the Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, the Laser Research Laboratory, Howe Laboratory of Ophthalmology, Massachusetts Eye and Ear Infirmary, Boston, the Wellman Laboratories, Massachusetts General Hospital, Boston, and Institut fur Biophysik, Universitat Frankfurt, Frankfurt, West Germany. Presented at the Eighty-ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15, 1984. Supported in part by a grant from the Donaldson Trust and PHS Grant No. 1 R23 EY05442-01. Reprint requests to Carmen A. Puliafito, MD, Howe Laboratory of Ophthalmology, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114.
producing tissue ablation with a high degree of precision and with minimal thermal damage to adjacent structures. We performed comparative experimental studies of excimer laser ablation of the cornea and crystalline lens using both 193 nm (ArF) and 248 nm (krypton fluoride, KrF) radiation. To characterize the biologic effects and damage mechanisms associated with excimer laser ablation, corneal tissue adjacent to the region of excimer laser ablation was examined using transmission and scanning electron microscopy. Lens tissue adjacent to the laser ablation was studied using transmission and scanning electron microscopy. Spectral absorption of the human cornea in the ultraviolet was studied.
MATERIALS AND METHODS CORNEA STUDIES
Ablation studies. The experimental apparatus is shown schematically in Figure 1. A Lambda Physik EMG 10 1 741
OPHTHALMOLOGY
Excimer Laser _____
•
JUNE 1985 •
1__"1-1ttt LI1+1---1.__ 0 1 -
~_.....rAPerture
Mask Attenuators c:....;:::J
Eye
Fig ). Schematic diagram of experimental apparatus for ablation studies.
excimer laser waS used to produce 193 (ArF) or 248 (KrF) nm radiation. The laser pulse width was 12 ns. Attenuators reduced the laser beam energy to desired levels. The attenuated beam was aimed through either a UV-grade cylindrical quartz lens with a focal length of 11.5 cm or a mask. The mask consisted of a chromeplated quartz flat in which slits 5, 10, 25, 50, 100 and 200 11m wide and 6 mm long had been etched by photolithography. Energy measurements were made with a joulemeter (Geritec ED-200) monitored by a storage oscilloscope (Tektronix 7633). Four freshly enucleated human eyes with normal corneas and six freshly enucleated calf eyes were used in the in vitro studies, When the cylindrical lens was used, an energy reading was taken 5 cm behind the lens (with the laser operating at a 2 Hz repetition rate) and the eye was placed in the focal plane for exposure. When the mask was used, the energy reading was taken just behind the mask and the eye was positioned as close to the mask as possible without contact. The corneal surface was oriented perpendicularly to the beam and exposed to a given number of pulses. The eyes were examined under a dissecting microscope immediately after exposure. The area of corneal exposure with the mask in place was assumed to be the area of the slit. The exposure area when the quartz lens was used was determined by measuring the imprint made by the laser beam on a piece of developed photographic print paper. In the initial set of exposures at 193 nm, two bovine and two human eyes were irradiated. Mask width, repetition rate (5-50 Hz), number of pulses (100-20,000), and energy per unit area (fluence) per pulse (approximately 90-140 mJ/cm 2/pulse) were varied in order to determine the experimental conditions which would yield substantial ablation. A total of 16 exposures were made. In the second set of in vitro exposures, in one human and two bovine eyes, mask width (100 11m), repetition rate (50 Hz), a number of pulses (10,000) were kept constant while the fluence was varied between 8 and 60 mJ/cm 2 per pulse. A total of 13 exposures were made. A similar study was performed at 248 nm using two bovine eyes and one human eye. The repetition rate was held constant at 50 Hz, the number of pulses ranged between 1000 and 10,000, and the fluence was varied between 30 and 940 mJ/cm 2 per pulse. A total of 13 exposures were made. Cuts were made with a diamond knife (Micra Titanium Corporation) in the cornea of two human eye bank eyes for comparison purposes. 742
VOLUME 92
•
NUMBER 6
Sample between
Excimer Laser
l_&_~"~r~~~::+_r\ _"'-."w T I! V·i" L-n-"
- - - - - ' Attenuato'.1
Apertures
Saline Filter
Lens
~
V
Vacuum Phatodiode
Fig 2. Schematic diagram of experimental apparatus for 193 nm corneal transmission study.
The human and bovine eyes were fixed immediately after treatment in a 2% gluteraldehyde, 4% paraformaldehyde, 0.1 M sodium cacodylate buffer solution. Corneas were later processed for light, transmission electron (TEM), and scanning electron (SEM) microscopy. For light microscopy, corneas were embedded in JB-4 glycol methacrylate, cut at 2 to 311m, and stained with Stevenol's blue. Corneas for TEM and SEM were postfixed in I %osmium tetroxide. TEM tissue was embedded in Spurr, cut at approximately 700 A, and stained with uranyl acetate and lead citrate. Specimens were examined using a JEOL 100C transmission electron microscope and a JEOL 35 scanning electron microscope. Spectral transmission studies. Sample preparation. Corneas were dissected from freshly enucleated calf and human eyes. A square measuring approximately 0.5 to 1.0 cm was cut from the center of each cornea. Each block of tissue was frozen to a specimen holder covered with OCT compound and cut on a freezing microtome at approximately -13°C, with nominal thickness settings at 32 11m. Samples were floated in physiological saline onto a 2.5 mm thick quartz flat and covered by another 2.5 mm thick quartz flat. The sandwich was then covered with a metal mask containing two holes, one covering the sample, the other serving as reference location. The apertures were 3, 4, or 6 mm in diameter, depending on the size of the specimen. Samples were inspected under a stereoscopic microscope for proper masking of the specimen and optical transparency. Samples were checked under phase contrast to ensure that no holes or tears were present. After mounting using the fine focus micrometer on a light microscope sample thicknesses were measured; the position of the front and back could be determined to a precision of 211m. These measurements revealed a range in the sample thickness of approximately 25%. Transmission measurements (193 nm). The transmission of the corneal sections at 193 nm was measured using the ArF laser. The laser measurement apparatus is schematically drawn in Figure 2. The Lambda-Physik EMG 101 laser was set to emit 100 mJ/pulse at 193 nm. The beam was attenuated such that the sample surface fluence was below 1 mJ/cm 2/pulse, well below ablation threshold. A lens with a focal length of 5 cm, placed 6 cm from the sample, imaged the sample onto a vacuum photodiode via a 45° dielectric mirror. Three possible measurement errors were carefully checked: stray light reaching the detector via a path other than through the specimen, fluorescence from
PULIAFITO, et al
Fig 3. Light micrograph of a 193 nm, trough-like ablation in a bovine eye. Dosage parameters: 300 pulses, 5 Hz, 134 mJ /cm 2/pulse, 200 ILm mask (original magnification X733).
optical components in the beam path, or from the sample and light, scattered in the sample. Stray light was minimized by enclosing the beam path between the 45° mirror and the photodetectors. An additional filter was prepared by filling a 3 em pathlength cuvette with physiological saline. This filter has an absorbance greater than 100 at 193 nm and essentially full transmission at 200 nm. This filter was used to check that none of the measured light was fluorescence from either the sample or optical components. Transmission of a He-Ne laser beam through mounted samples revealed substantial scattering with a strong forward characteristic. The aperture of the imaging lens was chosen large enough to collect most of this scattered light. Spectrophotometer study. Five specimens were transferred into a Beckmann 5270 spectrophotometer for measurement immediately after the laser measurements, remaining in the same amount between the quartz flats. Reference and sample beam were both apertured down to a 6 rom diameter area and a zero absorption line was obtained throughout the wavelength range of 700 to 200 nm with no sample in place. The Beckmann 5270 is unreliable for absorbances greater than 2 at wavelengths less than 230 nm. LENS ABLA nON STUDIES
A Questek series 2000 excimer laser provided 193 nm (15 nsec) pulses and 248 nm (20 nsec) pulses of radiation. The energy output was controlled via appropriate gas combinations and the laser microcomputer. A UV-grade cylindrical quartz lens with focal length of 20 cm was used to focus the beam. Crystalline lenses used in the study were removed from fresh calf eyes. The bovine lens was placed in the focal plane of the quartz lens. The procedure for ablation paralleled that for the cornea. The repetition rate for all lens work was 5 Hz. Eight bovine lenses were exposed to 100 to 400 pulses with fluences ranging from 1000 to 3800 mJ/cm 2 per pulse at 193 nm and 8 lenses were exposed to 75 to 300 pulses with fluences ranging from 2200 to 3800
• EXCIMER LASER
or
Fig 4. Light micrograph 193nm, slit-like ablation in a bovine eye. Dosage parameters: 10,000 pulses, 50 Hz, 93 mJ/cm 2/pulse, 25 ILm mask (original magnification X 165).
Fig S. Light micrograph of a 193 nm, slit-like ablation in a human cornea. Dosage parameters: 20,000 pulses, 50 Hz, 125 mJ/cm 2/pulse, 10 ILm mask (original magnification X214).
mJ/cm 2 per pulse at 248 nm. Specimens were prepared for TEM and SEM study as previously described for the cornea study.
RESULTS CORNEA STUDIES
Ablation study. The lowest fluences at which ablation was observed under 3X magnification were 46 mJ/cm 2 per pulse for a human eye and 20 mJ/cm 2 per pulse for a .bovine eye for 193 nm radiation and 58 mJ/cm 2 per pulse and 71 mJ/cm 2 per pulse for a human and bovine eye, respectively, at 248 nm. In general, a higher fluence was needed for the 248 nm radiation to produce results similar to 193 nm ablations. The variation of the appearance of the ablations was unrelated to repetition rate. The 193 nm ablations were of two types: trough-like and slit-like (Figs 3, 4, 5), depending on the width of 743
OPHTHALMOLOGY
•
JUNE 1985
•
VOLUME 92
•
NUMBER 6
Fig 6. Transmission electron micrograph of the edge of the corneal stromal ablation shown in Figure 5. A 0. 1 to 0.3 I'm-wide zone of abnormal tissue or debris (arrows) is demonstrated adjacent to structurally normal collagen on the wall of the laser incision (original magnification X10,700).
Fig 7. Higher power electron micrograph of the specimen shown in Figures 5 and 6. The zone of abnormality consists of an outer densely staining region (a), middle lightly staining region (b), and an inner region (c) showing some increased staining in an area where the fine structure of collagen is partially preserved. Each zone measures approximately 0.07 I'm (original magnification X66,Ooo).
Fig 8. Scanning electron micrograph showing sharp cleavage of the corneal epithelium (arrow) and stroma (arrowhead) in a 193 nm ablation . Dosage parameters: 20,000 pulses, 50 Hz, 125 mJ/cm 2/pulse, 10 I'm mask (original magnification X570).
Fig 9. Light micrograph of a 248 nm ablation. Note irregular edge and disorganization of stromal collagen adjacent to the ablation. Dosage parameters: 1050 pulses, 50 Hz, 190 mJ /cm 2/pulse, quartz lens (original magnification X61O).
the corneal exposure. The sides of the slit ablations were smooth. TEM reveals a zone of damaged stromal tissue approximately 0.1 to 0.3 tim thick on the edges of the ablation (Figs 6, 7) with preservation of corneal fine structures beyond this region. SEM study gives further evidence of the sharp demarcation of the cuts (Fig 8). Ablation with 248 nm radiation at comparable fluences produced incisions with ragged edges and with a much wider region of adjacent stromal damage at least 2.5 tim wide (Figs 9, 10). Not only is the zone of damage wider at 248 nm than at 193 nm (for comparable fluences), but the character of the tissue alterations are more severe, with marked disorganization and disruption of stromal collagen. The ablations made with the cylindrical quartz lens in place were similar to those made with the
mask. TEM of diamond knife cuts showed a very narrow band of densely staining material approximately 0.02 to 0.04 tim in width along the edge of the incisions, with preservation of normal structure in adjacent tissue (Fig 11). Spectral studies. Transmission measurements (193 nm). All eight samples of 32 tim nominal thickness gave very consistent results. Since human and bovine corneas had essentially the same absorption, we have combined their data to obtain an average absorbance, A, of 3.79 with a standard deviation of 0.24. The exponential law of absorption states that T = IO- A = e- ud where T is the transmission, a is the absorption coefficient, and d is the thickness, 32 X 10- 4 cm. Since ad = 2.3 X A = 8.72, the absorption coefficient a is approximately
744
PULIAFITO, et al •
Fig 10. Transmission electron micrograph of specimen shown in Figure 9. The corneal stroma adjacent to the region of ablation (A) shows a broad zone of disorganization measuring at least 2.S !Lm wide (original magnification X8580).
0,2 0L-2-0LO--L--2L40 --L--2~8-0~~3=2-0--
Wavelength (nm) Fig 12. Plot of absorbance versus wavelength in the far ultraviolet spectrum for a 32 !Lm thick section of bovine cornea. The point at 193 nm represents the average of laser-transmission measurements made on 8 different 32 !Lm thick samples. The error bar indicates one stand~rd deviation.
2700 em-I . The corresponding absorption length, a - I, at which 63% of the radiation is absorbed, is approximately 3.70 Jlm. Spectrophotometer study. A sample spectrum taken on a 32 Jlm bovine sample is shown in Figure 12. The average tissue absorption length at 248 nm, determined from the five spectra, is 47 ± 7 Jlm which corresponds to an absorption coefficient of 210 em- I. We estimate an uncertainty of 30% in absolute measurements, due mainly to the uncertainty of the thickness of the section. Lens study. Definite ablative effects were seen in all
EXCIMER LASER
Fig 11. Transmission electron micrograph of the edge of a diamondknife cut in a human cornea. A narrow band of densely staining material approx. 0.02 to 0.04 !Lm wide is seen at the edge of the cut (original magnification XSO,OOO).
Fig 13. Scanning electron micrograph of the region showing ablation of a bovine crystalIine lens. Laser light at 193 nm was focused with a cylindrical quartz lens and resulted in a region of ablated lens protein with smooth walIs (arrows). Linear cracks seen are artifacts of tissue processing. Dosage parameters: 400 pulses, 5 Hz, 3400 mJ/cm 2/pulse (original magnification X702).
but one lens exposure. A dose of 400 pulses of 1000 mJ/cm 2/pulse, 193 nm radiation, resulted in bubble formation beneath the capsule and no visible ablation under 3X magnification. Further investigation at lower fluences is necessary to determine the significance of this effect. All other exposures perforated the capsule and ablated lens protein. Figure 13 illustrates an ablated region with smooth walls and without evidence of charring or carbonization. The effects of' the 193 and 248 nm radiation were comparable on gross examination. Transmission electron micrography, however, showed more pronounced and deeper damage along the edge of the 248 nm ablation compared to 193 nm ablation (Figs 14, 15). 745
OPHTHALMOLOGY
•
JUNE 1985
Fig 14. Transmission electron micrograph of lens protein adjacent to the region of 193 nm ablation (A) showing a zone of disruption (8) adjacent to normal lens ultrastructure (C). Dosage parameters: 400 pulses,5 Hz, 1900 mJ/cm 2/pulse (original magnification X4940).
DISCUSSION Lasers have been used in a variety of surgical applications to remove tissue by producing local heating. The high power that lasers can produce, along with the fact that the output radiation can be focused down to a small spatial region, allows highly localized heating. Tissue is removed when the local temperature becomes high enough to cause vaporization of water in the tissue. The penetration of laser radiation into tissue varies with wavelength; at the CO 2 laser wavelength of 10.6 /-Lm the absorption length of the radiation in tissue, determined primarily by the absorption in water, is about 10 /-Lm.? Continuous wave argon lasers operating at 514.5 or 488 nm can be used for surface ablation of absorbing tissues as can be continuous wave neodymium:YAG laser at 1064 nm. Although laser radiation can be focused to spot sizes ranging from microns to tens of microns, depending on the laser wavelength and optics used, both diffusion of heat out of the focal region of the beam in the case of the CO 2 laser and relatively deep penetration of the radiation into tissue (in the case of the neodymium: YAG and argon lasers) can lead to regions of thermal damage that are substantially larger than the size of the focal spot. 8 Such diffusion of heat away from the focal region can coagulate adjacent blood vessels and lead to hemostasis which can be clinically useful in the case of vascularized tissues. Destruction of viable tissue can also be caused by these effects, however. The development of high-power pulsed lasers emitting in the ultraviolet (UV) region of the spectrum, specifically from 193 to 351 nm, has led to experiments to evaluate their usefulness in processing materials. 1.2.9. 10 In polymer targets, the penetration depth of 193 nm radiation is short, typically microns, leading to very different materials processing results than are achieved using more
746
•
VOLUME 92
•
NUMBER 6
Fig 15. Transmission electron micrograph of lens protein adjacent to the region of 248 nm ablation (A) showing increased disruption and variable staining of the adjacent tissue (8) compared to 193 nm. Normal ultrastructure is seen in region (C). Dosage parameters: 400 pulses, 5 Hz, 1800 mJ/cm 2/pulse (original magnification X4940).
traditional visible and infrared laser sources. Initial experiments with organic polymers showed that with appropriate optics submicrometer patterns could be formed by material removal; depth of the cuts could also be controlled precisely, with approximately 0. 1 /-Lm of polymer removed per laser pulse. I •10 In addition, the energy of the UV photons is high, up to 6.4 eV for 193 nm, opening the possibility of breaking chemical bonds in the target rather than simply heating it. Subsequently, a number of biologic targets including cornea, skin, and atherosclerotic lesions have been ablated using ultraviolet excimer laser radiation. 3-6 In our studies, we compared ablation of the cornea using 193 and 248 nm excimer radiation. We observed that the threshold fluence for corneal ablation was higher at 248 nm than at 193 nm. The most impressive ablation of corneal stroma was produced by 193 nm radiation. Using appropriately configured photolithographed masks, cuts as narrow as 20 /-Lm in width could be produced. As demonstrated by Trokel and co-workers, 3 we found the depth of the incision could be controlled by varying the fluence or number of pulse applications. In general, the higher the fluence, the greater the amount of material ablated per pulse. When examined by light microscopy, the corneal stroma adjacent to 193 nm excimer incisions showed no histologic evidence of thermal injury, even when the ablation was performed with fluences as high as 200 mJ/cm. When these specimens were examined by transmission electron microscopy, a narrow zone of damaged tissue (0.1-0.3 /-Lm) is seen immediately adjacent to the tissue which was removed by the laser. This edge effect may represent a thin band of thermal .denaturation or photoablated material which was not completely ejected and adhered to the wall of the incision. Beyond this narrow zone of apparent damage, the fine structure of stromal collagen is remarkably well preserved. This discrete zone of
PULIAFITO, et al
altered tissue was not seen in the case of corneal incisions made with a diamond knife, although a minimal edge effect was observed in those incisions as well. Scanning electron microscopy (SEM) of lesions made with 193 nm showed that the inside wall of the laser incision was smooth and the incision through the epithelium was sharp. Scanning electron microscopy of the 193 nm excimer cuts were similar to previously published observations of diamond knife cuts in the cornea. II The high degree of precision and minimal surrounding tissue damage observed after excimer laser ablation of the cornea with 193 nm radiation are in sharp contrast to the histologic findings reported after COz laser surgery of the cornea and sclera using either continuous wave or pulsed laser sources. COz laser ablation of the cornea and sclera typically produces marked thermal damage in adjoining tissue, with coagulative necrosis, charring and carbonization. lz- 16 Decreased thermal injury is associated with the use of shorter pulses (which reduce the effect of thermal diffusion to adjacent structures) but even use of short pulsed COz radiation has not produced results comparable to the excimer laser. 17 Incisions made with 248 nm radiation at fluences comparable to the 193 nm experiments showed greater damage to tissue adjacent to the area which had been ablated. These alterations included denaturation of stromal collagen with ultrastructural evidence of marked disorganization and alteration. Edges of incisions made with 248 nm were irregular, in contrast to the sharp edges seen at 193 nm. The reasons for greater tissue damage in the case of 248 nm ablation compared to 193 nm ablation are not established, but may be related to the greater penetration depth of 248 nm radiation in the cornea when compared to 193 nm radiation. In our spectral transmission studies, we found the absorption length of 248 nm radiation to be more than ten times greater then that of 193 nm radiation: 47 JIm for 248 nm, compared to 4 JIm for 193 nm. The excimer laser was also effective in producing rapid and well controlled ablation of the bovine crystalline lens protein in vitro. Scanning electron microscopy revealed a smooth walled crater of ablation, without evidence of charring or carbonization. Transmission electron microscopy revealed a narrow zone of tissue damage in the case of 193 nm laser ablation and a wider zone in the instance of 248 nm ablation, correlating with the findings in corneal stromal protein ablation. The relative involvement of photochemical and thermal processes in excimer laser ablation is currently being debated. Srinivasan and co-workers have termed the process "ablative photodecomposition" wherein molecules at the irradiated surface are broken into smaller volatile fragments without heating the remaining substrate. 1,2 It has been further proposed that the mechanism of action is "photochemical;" that is, high energy ultraviolet photons directly break molecular bonds, in contrast to ablation with laser wavelengths in the infrared region where absorption is due to vibrational transitions of molecules, producing an increase in temperature, with resultant thermal effects. 18
•
EXCIMER LASER
A number of substances in the cornea are possible chromophores of importance in excimer laser tissue ablation. At 248 nm, possible chromophores are nucleic acids, which, in the cornea, are found mainly in the epithelium. 19 ,2o All bases in nucleic acids absorb well between 230 and 290 nm (due to the 260 nm absorption peak of the purines and pyrimidines), approximately 10 to 20 times greater than proteins in the same spectral region. zl Nucleic acids absorb even more strongly at 193 nm than at 248 nm.zo Collagen, the protein which comprises approximately 70% of the dry weight of the stroma, has spectral absorption that begins to rise at wavelengths below 260 nm and increases rapidly below 240 nm,z2 corresponding to absorption by nonaromatic amino acids. At 193 nm, a primary chromophore may be the peptide bonds in collagen. The photon energy of 6.4 eV at 193 nm is more than sufficient to cleave peptide bonds (bond energy of 3 eV) or the adjacent carbon-carbon bond (bond energy of 3.5 eV) of the polypeptide chain. Another potential chromophore of importance may be the glycosaminoglycans, which are located in the interfibrillar and interstitial space of the stroma and comprise approximately 4.5% of its dry weight. 23 Absorption data for the glycosaminoglycans indicate absorption peaks around 190 nm, with no significant absorption at 248 nm. Z4 ,25 It is possible that the high quantum yield for peptide bond cleavage at 193 nm, coupled with the high absorption of glycosaminoglycans which are intertwined throughout the collagen moiety, is responsible for the superior tissue removal and lower fluence requirement associated with 193 nm ablation. The excimer laser may have potential for clinical application in corneal surgery. Techniques similar to those outlined in this paper may be used to make radial corneal incisions, as well as etching or lathing the cornea with a high degree of spatial resolution. Such technology would offer the possibility of non-contact corneal cutting, eliminating alterations in corneal topography produced by the surgical instrument. This may facilitate the use of real time topographical analysis systems to monitor or control the surgical process. The corneal stroma, which is composed primarily of extracellular matrix with few cells and with no blood vessels, would appear to be an excellent biologic target for UV ablation. Use of the excimer laser for cutting or removing tissue from inside the eye presents a number of technical challenges. Ultraviolet laser tissue ablation occurs at megawatt peak powers; such high peak powers cannot be transmitted with current optical fiber technology. Additionally, chloride ions in physiologic saline are highly absorbing at 193 nm and may block the ablation process. 4 Nevertheless, this technology may not only be useful in lens surgery, but offers, in principle, the promise of cutting vitreous or epiretinal membranes with a high degree of precision. Finally, clinical acceptance of ultraviolet laser ablation will require an evaluation of potential UV-induced carcinogenesis and mutagenesis. UV radiation is known to cause the formation of DNA photoproducts, mutation, 747
OPHTHALMOLOGY
•
JUNE 1985 •
transformation, and cell death. 26 Experimental studies to investigate potential carcinogenesis following laser ablation of the cornea are important.
13. 14.
REFERENCES 1. Srinivasan R, Mayne-Banton V. Self-developing photoetching of poly(ethylene terephthalate) films by far-ultraviolet excimer laser radiation. Appl Phys Lett 1983; 41:576-8. 2. Srinivasan R, Leigh WJ. Ablative photodecomposition: action of farultraviolet (193 nm) laser radiation on poly(ethylene terephthalate) films. J Am Chem Soc 1982; 104:6784-5. 3. Trokel SL, Srinivasan R, Braren B. Excimer laser surgery of the comea. Am J Ophthalmol1983; 96:710-5. 4. Lane RJ, Linsker R, Wynne JJ. Ultraviolet-laser ablation of skin and other tissue. Paper FL-1. Technical Digest, Conference on Laser and Electro-optics. Anaheim, Califomia, June 19-22, 1984. 5. Linsker R, Srinivasan R, Wynne JJ, Alonso DR. Far-ultraviolet laser ablation of atherosclerotic lesions. Lasers Surg Med 1984; 4:201-6. 6. Grundfest WS, Litvack IF, Morgenstem L, Forrester JS. Effect of laser irradiation on human atherosclerotic aorta: amelioration of laser-induced thermal damage. Paper FL-2. Technical Digest, Conference on Laser and Electro-Optics, Anaheim, CA, June 19-22, 1984. 7. Mainster MA. Ophthalmic applications of infrared lasers-thermal considerations. Invest Ophthalmol Vis Sci 1979; 18:414-20. 8. Cummins L, Nauenberg M. Thermal effects of laser radiation in biological tissue. Biophys J 1983; 42:99-102. 9. Koren G, Yeh JTC. Emission spectra, surface quality, and mechanism of excimer laser etching of polyimide films. Appl Phys Lett 1984; 44:1112-4. 10. Deutsch TF, Geis MW. Self-developing UV photoresist using excimer laser exposure. J Appl Phys 1983; 54:7201-4. 11. Galbavy EJ. Use of diamond knives in ocular surgery. Ophthalmic Surg 1984; 15:203-5. 12. Beckman H, Rota A, Barraco R, et al. Limbectomies, keratectomies,
748
15.
16.
17.
18.
19.
20. 21.
22. 23. 24.
25. 26.
VOLUME 92 •
NUMBER 6
and keratostomies performed with a rapid-pulsed carbon dioxide laser. Am J Ophthalmol 1971; 71: 1277 -83. Leibowitz HM, Peacock GR. Comeal injury produced by carbon dioxide laser radiation. Arch Ophthalmol1969; 81:713-21. Keates RH, Pedrotti LS, Weichel H, Possel WHo Carbon dioxide laser beam control for comeal surgery. Ophthalmic Surg 1981; 12: 117-22. Kilp H, Walzer K, Tackel M, Horster B. Some experiments on extraocular and intraocular treatment with CO2 lasers. Doc Ophthalmol Proc Ser 1984; 36:265-70. Fine BS, Fine S, Peacock GR, et al. Preliminary observations on ocular effects of high-power, continuous CO2 laser irradiation. Am J Ophthalmol 1967; 64:209-22. Miller JB, Smith MR. Current status of pars plana endophotocoagulation and endophotoincision using carbon dioxide, argon, and krypton laser. Doc Ophthalmol Proc Ser 1984; 36:235-58. Srinivasan R, Braren-Bodil. Ablative photodecomposition of polymer films by pulsed far-ultraviolet (193 nm) laser radiation: Dependence of etch depth on experimental conditions. J Polym Sci Polym Chem Ed 1984; 22:2601-9. Taboada J, Mikesell GW Jr, Reed RD. Response of the comeal epithelium to KrF excimer laser pulses. Health Phys 1981; 40:67783. O'Brien WJ. Measurement of comeal DNA content. Invest Ophthalmol Vis Sci 1979; 18:538-43. Patrick MH. Physical and chemical properties of DNA. In: Wang SY, ed. Photochemistry and Photobiology of Nucleic Acids. New York: Academic Press, 1976; Vol 2:22. Loofbourow JR, Gould BS, Sizer IW. Studies of the ultraviolet absorption spectra of collagen. Arch Biochem 1949; 22:406-11. Maurice DM. The comea and sclera. In: Davson H, ed. The Eye, 3d ed. New York: Academic Press, 1984; Vol. 1b:12. Stone AL. Optical rotary dispersion of mucopolysaccharides. III. Ultraviolet circular dichroism and conformational specificity in amide groups. Biopolymers 1971; 10:739-51. Stone AL. Personal communication. Epstein WL, Fukuyama K, Epstein JH. Early effects of ultraviolet light on DNA synthesis in human skin in vivo. Arch Dermatol 1969; 100:84-9.
Abstract: The pulsed ultraviolet excimer laser has been used to produce tissue ablation with a high degree of precision and with minimal thermal damage to adjacent structures. In comparative studies of excimer laser ablation of the cornea and crystalline lens using 193 nm and 248 nm radiation, threshold fluence for corneal and lens ablation was higher at 248 nm than at 193 nm. Ablation of corneal stroma at 193 nm produced the most precise cuts. When examined by transmission electron microscopy, a narrow zone of damaged tissue (0.1 to 0.3 Jlm) was seen immediately adjacent to the tissue removed by the laser. Ablation with 248 nm radiation produced incisions with ragged edges and with a wider and more severe zone of damage in adjacent stroma. Ultraviolet spectral transmission studies of the corneal stroma showed that absorption is 10 times greater at 193 nm than at 248 nm. The excimer laser was effective in producing well controlled ablation of the crystalline lens in vitro, with effects parallel to those seen in the cornea. [Key words: cornea, excimer, keratorefractive surgery, laser, lens.] Ophthalmology 92:741-748, 1985
Excimer lasers are high-power sources of pulsed ultraviolet radiation. Recently, 193 nm radiation from an argon-fluoride (ArF) excimer laser has been shown to produce precise etching of polymer films and clean incisions in biologic materials, in particular the cornea, skin and aorta. I - 6 The excimer laser has been proposed as a new laser therapeutic device with the potential of
From the Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, the Laser Research Laboratory, Howe Laboratory of Ophthalmology, Massachusetts Eye and Ear Infirmary, Boston, the Wellman Laboratories, Massachusetts General Hospital, Boston, and Institut fur Biophysik, Universitat Frankfurt, Frankfurt, West Germany. Presented at the Eighty-ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15, 1984. Supported in part by a grant from the Donaldson Trust and PHS Grant No. 1 R23 EY05442-01. Reprint requests to Carmen A. Puliafito, MD, Howe Laboratory of Ophthalmology, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114.
producing tissue ablation with a high degree of precision and with minimal thermal damage to adjacent structures. We performed comparative experimental studies of excimer laser ablation of the cornea and crystalline lens using both 193 nm (ArF) and 248 nm (krypton fluoride, KrF) radiation. To characterize the biologic effects and damage mechanisms associated with excimer laser ablation, corneal tissue adjacent to the region of excimer laser ablation was examined using transmission and scanning electron microscopy. Lens tissue adjacent to the laser ablation was studied using transmission and scanning electron microscopy. Spectral absorption of the human cornea in the ultraviolet was studied.
MATERIALS AND METHODS CORNEA STUDIES
Ablation studies. The experimental apparatus is shown schematically in Figure 1. A Lambda Physik EMG 10 1 741
OPHTHALMOLOGY
Excimer Laser _____
•
JUNE 1985 •
1__"1-1ttt LI1+1---1.__ 0 1 -
~_.....rAPerture
Mask Attenuators c:....;:::J
Eye
Fig ). Schematic diagram of experimental apparatus for ablation studies.
excimer laser waS used to produce 193 (ArF) or 248 (KrF) nm radiation. The laser pulse width was 12 ns. Attenuators reduced the laser beam energy to desired levels. The attenuated beam was aimed through either a UV-grade cylindrical quartz lens with a focal length of 11.5 cm or a mask. The mask consisted of a chromeplated quartz flat in which slits 5, 10, 25, 50, 100 and 200 11m wide and 6 mm long had been etched by photolithography. Energy measurements were made with a joulemeter (Geritec ED-200) monitored by a storage oscilloscope (Tektronix 7633). Four freshly enucleated human eyes with normal corneas and six freshly enucleated calf eyes were used in the in vitro studies, When the cylindrical lens was used, an energy reading was taken 5 cm behind the lens (with the laser operating at a 2 Hz repetition rate) and the eye was placed in the focal plane for exposure. When the mask was used, the energy reading was taken just behind the mask and the eye was positioned as close to the mask as possible without contact. The corneal surface was oriented perpendicularly to the beam and exposed to a given number of pulses. The eyes were examined under a dissecting microscope immediately after exposure. The area of corneal exposure with the mask in place was assumed to be the area of the slit. The exposure area when the quartz lens was used was determined by measuring the imprint made by the laser beam on a piece of developed photographic print paper. In the initial set of exposures at 193 nm, two bovine and two human eyes were irradiated. Mask width, repetition rate (5-50 Hz), number of pulses (100-20,000), and energy per unit area (fluence) per pulse (approximately 90-140 mJ/cm 2/pulse) were varied in order to determine the experimental conditions which would yield substantial ablation. A total of 16 exposures were made. In the second set of in vitro exposures, in one human and two bovine eyes, mask width (100 11m), repetition rate (50 Hz), a number of pulses (10,000) were kept constant while the fluence was varied between 8 and 60 mJ/cm 2 per pulse. A total of 13 exposures were made. A similar study was performed at 248 nm using two bovine eyes and one human eye. The repetition rate was held constant at 50 Hz, the number of pulses ranged between 1000 and 10,000, and the fluence was varied between 30 and 940 mJ/cm 2 per pulse. A total of 13 exposures were made. Cuts were made with a diamond knife (Micra Titanium Corporation) in the cornea of two human eye bank eyes for comparison purposes. 742
VOLUME 92
•
NUMBER 6
Sample between
Excimer Laser
l_&_~"~r~~~::+_r\ _"'-."w T I! V·i" L-n-"
- - - - - ' Attenuato'.1
Apertures
Saline Filter
Lens
~
V
Vacuum Phatodiode
Fig 2. Schematic diagram of experimental apparatus for 193 nm corneal transmission study.
The human and bovine eyes were fixed immediately after treatment in a 2% gluteraldehyde, 4% paraformaldehyde, 0.1 M sodium cacodylate buffer solution. Corneas were later processed for light, transmission electron (TEM), and scanning electron (SEM) microscopy. For light microscopy, corneas were embedded in JB-4 glycol methacrylate, cut at 2 to 311m, and stained with Stevenol's blue. Corneas for TEM and SEM were postfixed in I %osmium tetroxide. TEM tissue was embedded in Spurr, cut at approximately 700 A, and stained with uranyl acetate and lead citrate. Specimens were examined using a JEOL 100C transmission electron microscope and a JEOL 35 scanning electron microscope. Spectral transmission studies. Sample preparation. Corneas were dissected from freshly enucleated calf and human eyes. A square measuring approximately 0.5 to 1.0 cm was cut from the center of each cornea. Each block of tissue was frozen to a specimen holder covered with OCT compound and cut on a freezing microtome at approximately -13°C, with nominal thickness settings at 32 11m. Samples were floated in physiological saline onto a 2.5 mm thick quartz flat and covered by another 2.5 mm thick quartz flat. The sandwich was then covered with a metal mask containing two holes, one covering the sample, the other serving as reference location. The apertures were 3, 4, or 6 mm in diameter, depending on the size of the specimen. Samples were inspected under a stereoscopic microscope for proper masking of the specimen and optical transparency. Samples were checked under phase contrast to ensure that no holes or tears were present. After mounting using the fine focus micrometer on a light microscope sample thicknesses were measured; the position of the front and back could be determined to a precision of 211m. These measurements revealed a range in the sample thickness of approximately 25%. Transmission measurements (193 nm). The transmission of the corneal sections at 193 nm was measured using the ArF laser. The laser measurement apparatus is schematically drawn in Figure 2. The Lambda-Physik EMG 101 laser was set to emit 100 mJ/pulse at 193 nm. The beam was attenuated such that the sample surface fluence was below 1 mJ/cm 2/pulse, well below ablation threshold. A lens with a focal length of 5 cm, placed 6 cm from the sample, imaged the sample onto a vacuum photodiode via a 45° dielectric mirror. Three possible measurement errors were carefully checked: stray light reaching the detector via a path other than through the specimen, fluorescence from
PULIAFITO, et al
Fig 3. Light micrograph of a 193 nm, trough-like ablation in a bovine eye. Dosage parameters: 300 pulses, 5 Hz, 134 mJ /cm 2/pulse, 200 ILm mask (original magnification X733).
optical components in the beam path, or from the sample and light, scattered in the sample. Stray light was minimized by enclosing the beam path between the 45° mirror and the photodetectors. An additional filter was prepared by filling a 3 em pathlength cuvette with physiological saline. This filter has an absorbance greater than 100 at 193 nm and essentially full transmission at 200 nm. This filter was used to check that none of the measured light was fluorescence from either the sample or optical components. Transmission of a He-Ne laser beam through mounted samples revealed substantial scattering with a strong forward characteristic. The aperture of the imaging lens was chosen large enough to collect most of this scattered light. Spectrophotometer study. Five specimens were transferred into a Beckmann 5270 spectrophotometer for measurement immediately after the laser measurements, remaining in the same amount between the quartz flats. Reference and sample beam were both apertured down to a 6 rom diameter area and a zero absorption line was obtained throughout the wavelength range of 700 to 200 nm with no sample in place. The Beckmann 5270 is unreliable for absorbances greater than 2 at wavelengths less than 230 nm. LENS ABLA nON STUDIES
A Questek series 2000 excimer laser provided 193 nm (15 nsec) pulses and 248 nm (20 nsec) pulses of radiation. The energy output was controlled via appropriate gas combinations and the laser microcomputer. A UV-grade cylindrical quartz lens with focal length of 20 cm was used to focus the beam. Crystalline lenses used in the study were removed from fresh calf eyes. The bovine lens was placed in the focal plane of the quartz lens. The procedure for ablation paralleled that for the cornea. The repetition rate for all lens work was 5 Hz. Eight bovine lenses were exposed to 100 to 400 pulses with fluences ranging from 1000 to 3800 mJ/cm 2 per pulse at 193 nm and 8 lenses were exposed to 75 to 300 pulses with fluences ranging from 2200 to 3800
• EXCIMER LASER
or
Fig 4. Light micrograph 193nm, slit-like ablation in a bovine eye. Dosage parameters: 10,000 pulses, 50 Hz, 93 mJ/cm 2/pulse, 25 ILm mask (original magnification X 165).
Fig S. Light micrograph of a 193 nm, slit-like ablation in a human cornea. Dosage parameters: 20,000 pulses, 50 Hz, 125 mJ/cm 2/pulse, 10 ILm mask (original magnification X214).
mJ/cm 2 per pulse at 248 nm. Specimens were prepared for TEM and SEM study as previously described for the cornea study.
RESULTS CORNEA STUDIES
Ablation study. The lowest fluences at which ablation was observed under 3X magnification were 46 mJ/cm 2 per pulse for a human eye and 20 mJ/cm 2 per pulse for a .bovine eye for 193 nm radiation and 58 mJ/cm 2 per pulse and 71 mJ/cm 2 per pulse for a human and bovine eye, respectively, at 248 nm. In general, a higher fluence was needed for the 248 nm radiation to produce results similar to 193 nm ablations. The variation of the appearance of the ablations was unrelated to repetition rate. The 193 nm ablations were of two types: trough-like and slit-like (Figs 3, 4, 5), depending on the width of 743
OPHTHALMOLOGY
•
JUNE 1985
•
VOLUME 92
•
NUMBER 6
Fig 6. Transmission electron micrograph of the edge of the corneal stromal ablation shown in Figure 5. A 0. 1 to 0.3 I'm-wide zone of abnormal tissue or debris (arrows) is demonstrated adjacent to structurally normal collagen on the wall of the laser incision (original magnification X10,700).
Fig 7. Higher power electron micrograph of the specimen shown in Figures 5 and 6. The zone of abnormality consists of an outer densely staining region (a), middle lightly staining region (b), and an inner region (c) showing some increased staining in an area where the fine structure of collagen is partially preserved. Each zone measures approximately 0.07 I'm (original magnification X66,Ooo).
Fig 8. Scanning electron micrograph showing sharp cleavage of the corneal epithelium (arrow) and stroma (arrowhead) in a 193 nm ablation . Dosage parameters: 20,000 pulses, 50 Hz, 125 mJ/cm 2/pulse, 10 I'm mask (original magnification X570).
Fig 9. Light micrograph of a 248 nm ablation. Note irregular edge and disorganization of stromal collagen adjacent to the ablation. Dosage parameters: 1050 pulses, 50 Hz, 190 mJ /cm 2/pulse, quartz lens (original magnification X61O).
the corneal exposure. The sides of the slit ablations were smooth. TEM reveals a zone of damaged stromal tissue approximately 0.1 to 0.3 tim thick on the edges of the ablation (Figs 6, 7) with preservation of corneal fine structures beyond this region. SEM study gives further evidence of the sharp demarcation of the cuts (Fig 8). Ablation with 248 nm radiation at comparable fluences produced incisions with ragged edges and with a much wider region of adjacent stromal damage at least 2.5 tim wide (Figs 9, 10). Not only is the zone of damage wider at 248 nm than at 193 nm (for comparable fluences), but the character of the tissue alterations are more severe, with marked disorganization and disruption of stromal collagen. The ablations made with the cylindrical quartz lens in place were similar to those made with the
mask. TEM of diamond knife cuts showed a very narrow band of densely staining material approximately 0.02 to 0.04 tim in width along the edge of the incisions, with preservation of normal structure in adjacent tissue (Fig 11). Spectral studies. Transmission measurements (193 nm). All eight samples of 32 tim nominal thickness gave very consistent results. Since human and bovine corneas had essentially the same absorption, we have combined their data to obtain an average absorbance, A, of 3.79 with a standard deviation of 0.24. The exponential law of absorption states that T = IO- A = e- ud where T is the transmission, a is the absorption coefficient, and d is the thickness, 32 X 10- 4 cm. Since ad = 2.3 X A = 8.72, the absorption coefficient a is approximately
744
PULIAFITO, et al •
Fig 10. Transmission electron micrograph of specimen shown in Figure 9. The corneal stroma adjacent to the region of ablation (A) shows a broad zone of disorganization measuring at least 2.S !Lm wide (original magnification X8580).
0,2 0L-2-0LO--L--2L40 --L--2~8-0~~3=2-0--
Wavelength (nm) Fig 12. Plot of absorbance versus wavelength in the far ultraviolet spectrum for a 32 !Lm thick section of bovine cornea. The point at 193 nm represents the average of laser-transmission measurements made on 8 different 32 !Lm thick samples. The error bar indicates one stand~rd deviation.
2700 em-I . The corresponding absorption length, a - I, at which 63% of the radiation is absorbed, is approximately 3.70 Jlm. Spectrophotometer study. A sample spectrum taken on a 32 Jlm bovine sample is shown in Figure 12. The average tissue absorption length at 248 nm, determined from the five spectra, is 47 ± 7 Jlm which corresponds to an absorption coefficient of 210 em- I. We estimate an uncertainty of 30% in absolute measurements, due mainly to the uncertainty of the thickness of the section. Lens study. Definite ablative effects were seen in all
EXCIMER LASER
Fig 11. Transmission electron micrograph of the edge of a diamondknife cut in a human cornea. A narrow band of densely staining material approx. 0.02 to 0.04 !Lm wide is seen at the edge of the cut (original magnification XSO,OOO).
Fig 13. Scanning electron micrograph of the region showing ablation of a bovine crystalIine lens. Laser light at 193 nm was focused with a cylindrical quartz lens and resulted in a region of ablated lens protein with smooth walIs (arrows). Linear cracks seen are artifacts of tissue processing. Dosage parameters: 400 pulses, 5 Hz, 3400 mJ/cm 2/pulse (original magnification X702).
but one lens exposure. A dose of 400 pulses of 1000 mJ/cm 2/pulse, 193 nm radiation, resulted in bubble formation beneath the capsule and no visible ablation under 3X magnification. Further investigation at lower fluences is necessary to determine the significance of this effect. All other exposures perforated the capsule and ablated lens protein. Figure 13 illustrates an ablated region with smooth walls and without evidence of charring or carbonization. The effects of' the 193 and 248 nm radiation were comparable on gross examination. Transmission electron micrography, however, showed more pronounced and deeper damage along the edge of the 248 nm ablation compared to 193 nm ablation (Figs 14, 15). 745
OPHTHALMOLOGY
•
JUNE 1985
Fig 14. Transmission electron micrograph of lens protein adjacent to the region of 193 nm ablation (A) showing a zone of disruption (8) adjacent to normal lens ultrastructure (C). Dosage parameters: 400 pulses,5 Hz, 1900 mJ/cm 2/pulse (original magnification X4940).
DISCUSSION Lasers have been used in a variety of surgical applications to remove tissue by producing local heating. The high power that lasers can produce, along with the fact that the output radiation can be focused down to a small spatial region, allows highly localized heating. Tissue is removed when the local temperature becomes high enough to cause vaporization of water in the tissue. The penetration of laser radiation into tissue varies with wavelength; at the CO 2 laser wavelength of 10.6 /-Lm the absorption length of the radiation in tissue, determined primarily by the absorption in water, is about 10 /-Lm.? Continuous wave argon lasers operating at 514.5 or 488 nm can be used for surface ablation of absorbing tissues as can be continuous wave neodymium:YAG laser at 1064 nm. Although laser radiation can be focused to spot sizes ranging from microns to tens of microns, depending on the laser wavelength and optics used, both diffusion of heat out of the focal region of the beam in the case of the CO 2 laser and relatively deep penetration of the radiation into tissue (in the case of the neodymium: YAG and argon lasers) can lead to regions of thermal damage that are substantially larger than the size of the focal spot. 8 Such diffusion of heat away from the focal region can coagulate adjacent blood vessels and lead to hemostasis which can be clinically useful in the case of vascularized tissues. Destruction of viable tissue can also be caused by these effects, however. The development of high-power pulsed lasers emitting in the ultraviolet (UV) region of the spectrum, specifically from 193 to 351 nm, has led to experiments to evaluate their usefulness in processing materials. 1.2.9. 10 In polymer targets, the penetration depth of 193 nm radiation is short, typically microns, leading to very different materials processing results than are achieved using more
746
•
VOLUME 92
•
NUMBER 6
Fig 15. Transmission electron micrograph of lens protein adjacent to the region of 248 nm ablation (A) showing increased disruption and variable staining of the adjacent tissue (8) compared to 193 nm. Normal ultrastructure is seen in region (C). Dosage parameters: 400 pulses, 5 Hz, 1800 mJ/cm 2/pulse (original magnification X4940).
traditional visible and infrared laser sources. Initial experiments with organic polymers showed that with appropriate optics submicrometer patterns could be formed by material removal; depth of the cuts could also be controlled precisely, with approximately 0. 1 /-Lm of polymer removed per laser pulse. I •10 In addition, the energy of the UV photons is high, up to 6.4 eV for 193 nm, opening the possibility of breaking chemical bonds in the target rather than simply heating it. Subsequently, a number of biologic targets including cornea, skin, and atherosclerotic lesions have been ablated using ultraviolet excimer laser radiation. 3-6 In our studies, we compared ablation of the cornea using 193 and 248 nm excimer radiation. We observed that the threshold fluence for corneal ablation was higher at 248 nm than at 193 nm. The most impressive ablation of corneal stroma was produced by 193 nm radiation. Using appropriately configured photolithographed masks, cuts as narrow as 20 /-Lm in width could be produced. As demonstrated by Trokel and co-workers, 3 we found the depth of the incision could be controlled by varying the fluence or number of pulse applications. In general, the higher the fluence, the greater the amount of material ablated per pulse. When examined by light microscopy, the corneal stroma adjacent to 193 nm excimer incisions showed no histologic evidence of thermal injury, even when the ablation was performed with fluences as high as 200 mJ/cm. When these specimens were examined by transmission electron microscopy, a narrow zone of damaged tissue (0.1-0.3 /-Lm) is seen immediately adjacent to the tissue which was removed by the laser. This edge effect may represent a thin band of thermal .denaturation or photoablated material which was not completely ejected and adhered to the wall of the incision. Beyond this narrow zone of apparent damage, the fine structure of stromal collagen is remarkably well preserved. This discrete zone of
PULIAFITO, et al
altered tissue was not seen in the case of corneal incisions made with a diamond knife, although a minimal edge effect was observed in those incisions as well. Scanning electron microscopy (SEM) of lesions made with 193 nm showed that the inside wall of the laser incision was smooth and the incision through the epithelium was sharp. Scanning electron microscopy of the 193 nm excimer cuts were similar to previously published observations of diamond knife cuts in the cornea. II The high degree of precision and minimal surrounding tissue damage observed after excimer laser ablation of the cornea with 193 nm radiation are in sharp contrast to the histologic findings reported after COz laser surgery of the cornea and sclera using either continuous wave or pulsed laser sources. COz laser ablation of the cornea and sclera typically produces marked thermal damage in adjoining tissue, with coagulative necrosis, charring and carbonization. lz- 16 Decreased thermal injury is associated with the use of shorter pulses (which reduce the effect of thermal diffusion to adjacent structures) but even use of short pulsed COz radiation has not produced results comparable to the excimer laser. 17 Incisions made with 248 nm radiation at fluences comparable to the 193 nm experiments showed greater damage to tissue adjacent to the area which had been ablated. These alterations included denaturation of stromal collagen with ultrastructural evidence of marked disorganization and alteration. Edges of incisions made with 248 nm were irregular, in contrast to the sharp edges seen at 193 nm. The reasons for greater tissue damage in the case of 248 nm ablation compared to 193 nm ablation are not established, but may be related to the greater penetration depth of 248 nm radiation in the cornea when compared to 193 nm radiation. In our spectral transmission studies, we found the absorption length of 248 nm radiation to be more than ten times greater then that of 193 nm radiation: 47 JIm for 248 nm, compared to 4 JIm for 193 nm. The excimer laser was also effective in producing rapid and well controlled ablation of the bovine crystalline lens protein in vitro. Scanning electron microscopy revealed a smooth walled crater of ablation, without evidence of charring or carbonization. Transmission electron microscopy revealed a narrow zone of tissue damage in the case of 193 nm laser ablation and a wider zone in the instance of 248 nm ablation, correlating with the findings in corneal stromal protein ablation. The relative involvement of photochemical and thermal processes in excimer laser ablation is currently being debated. Srinivasan and co-workers have termed the process "ablative photodecomposition" wherein molecules at the irradiated surface are broken into smaller volatile fragments without heating the remaining substrate. 1,2 It has been further proposed that the mechanism of action is "photochemical;" that is, high energy ultraviolet photons directly break molecular bonds, in contrast to ablation with laser wavelengths in the infrared region where absorption is due to vibrational transitions of molecules, producing an increase in temperature, with resultant thermal effects. 18
•
EXCIMER LASER
A number of substances in the cornea are possible chromophores of importance in excimer laser tissue ablation. At 248 nm, possible chromophores are nucleic acids, which, in the cornea, are found mainly in the epithelium. 19 ,2o All bases in nucleic acids absorb well between 230 and 290 nm (due to the 260 nm absorption peak of the purines and pyrimidines), approximately 10 to 20 times greater than proteins in the same spectral region. zl Nucleic acids absorb even more strongly at 193 nm than at 248 nm.zo Collagen, the protein which comprises approximately 70% of the dry weight of the stroma, has spectral absorption that begins to rise at wavelengths below 260 nm and increases rapidly below 240 nm,z2 corresponding to absorption by nonaromatic amino acids. At 193 nm, a primary chromophore may be the peptide bonds in collagen. The photon energy of 6.4 eV at 193 nm is more than sufficient to cleave peptide bonds (bond energy of 3 eV) or the adjacent carbon-carbon bond (bond energy of 3.5 eV) of the polypeptide chain. Another potential chromophore of importance may be the glycosaminoglycans, which are located in the interfibrillar and interstitial space of the stroma and comprise approximately 4.5% of its dry weight. 23 Absorption data for the glycosaminoglycans indicate absorption peaks around 190 nm, with no significant absorption at 248 nm. Z4 ,25 It is possible that the high quantum yield for peptide bond cleavage at 193 nm, coupled with the high absorption of glycosaminoglycans which are intertwined throughout the collagen moiety, is responsible for the superior tissue removal and lower fluence requirement associated with 193 nm ablation. The excimer laser may have potential for clinical application in corneal surgery. Techniques similar to those outlined in this paper may be used to make radial corneal incisions, as well as etching or lathing the cornea with a high degree of spatial resolution. Such technology would offer the possibility of non-contact corneal cutting, eliminating alterations in corneal topography produced by the surgical instrument. This may facilitate the use of real time topographical analysis systems to monitor or control the surgical process. The corneal stroma, which is composed primarily of extracellular matrix with few cells and with no blood vessels, would appear to be an excellent biologic target for UV ablation. Use of the excimer laser for cutting or removing tissue from inside the eye presents a number of technical challenges. Ultraviolet laser tissue ablation occurs at megawatt peak powers; such high peak powers cannot be transmitted with current optical fiber technology. Additionally, chloride ions in physiologic saline are highly absorbing at 193 nm and may block the ablation process. 4 Nevertheless, this technology may not only be useful in lens surgery, but offers, in principle, the promise of cutting vitreous or epiretinal membranes with a high degree of precision. Finally, clinical acceptance of ultraviolet laser ablation will require an evaluation of potential UV-induced carcinogenesis and mutagenesis. UV radiation is known to cause the formation of DNA photoproducts, mutation, 747
OPHTHALMOLOGY
•
JUNE 1985 •
transformation, and cell death. 26 Experimental studies to investigate potential carcinogenesis following laser ablation of the cornea are important.
13. 14.
REFERENCES 1. Srinivasan R, Mayne-Banton V. Self-developing photoetching of poly(ethylene terephthalate) films by far-ultraviolet excimer laser radiation. Appl Phys Lett 1983; 41:576-8. 2. Srinivasan R, Leigh WJ. Ablative photodecomposition: action of farultraviolet (193 nm) laser radiation on poly(ethylene terephthalate) films. J Am Chem Soc 1982; 104:6784-5. 3. Trokel SL, Srinivasan R, Braren B. Excimer laser surgery of the comea. Am J Ophthalmol1983; 96:710-5. 4. Lane RJ, Linsker R, Wynne JJ. Ultraviolet-laser ablation of skin and other tissue. Paper FL-1. Technical Digest, Conference on Laser and Electro-optics. Anaheim, Califomia, June 19-22, 1984. 5. Linsker R, Srinivasan R, Wynne JJ, Alonso DR. Far-ultraviolet laser ablation of atherosclerotic lesions. Lasers Surg Med 1984; 4:201-6. 6. Grundfest WS, Litvack IF, Morgenstem L, Forrester JS. Effect of laser irradiation on human atherosclerotic aorta: amelioration of laser-induced thermal damage. Paper FL-2. Technical Digest, Conference on Laser and Electro-Optics, Anaheim, CA, June 19-22, 1984. 7. Mainster MA. Ophthalmic applications of infrared lasers-thermal considerations. Invest Ophthalmol Vis Sci 1979; 18:414-20. 8. Cummins L, Nauenberg M. Thermal effects of laser radiation in biological tissue. Biophys J 1983; 42:99-102. 9. Koren G, Yeh JTC. Emission spectra, surface quality, and mechanism of excimer laser etching of polyimide films. Appl Phys Lett 1984; 44:1112-4. 10. Deutsch TF, Geis MW. Self-developing UV photoresist using excimer laser exposure. J Appl Phys 1983; 54:7201-4. 11. Galbavy EJ. Use of diamond knives in ocular surgery. Ophthalmic Surg 1984; 15:203-5. 12. Beckman H, Rota A, Barraco R, et al. Limbectomies, keratectomies,
748
15.
16.
17.
18.
19.
20. 21.
22. 23. 24.
25. 26.
VOLUME 92 •
NUMBER 6
and keratostomies performed with a rapid-pulsed carbon dioxide laser. Am J Ophthalmol 1971; 71: 1277 -83. Leibowitz HM, Peacock GR. Comeal injury produced by carbon dioxide laser radiation. Arch Ophthalmol1969; 81:713-21. Keates RH, Pedrotti LS, Weichel H, Possel WHo Carbon dioxide laser beam control for comeal surgery. Ophthalmic Surg 1981; 12: 117-22. Kilp H, Walzer K, Tackel M, Horster B. Some experiments on extraocular and intraocular treatment with CO2 lasers. Doc Ophthalmol Proc Ser 1984; 36:265-70. Fine BS, Fine S, Peacock GR, et al. Preliminary observations on ocular effects of high-power, continuous CO2 laser irradiation. Am J Ophthalmol 1967; 64:209-22. Miller JB, Smith MR. Current status of pars plana endophotocoagulation and endophotoincision using carbon dioxide, argon, and krypton laser. Doc Ophthalmol Proc Ser 1984; 36:235-58. Srinivasan R, Braren-Bodil. Ablative photodecomposition of polymer films by pulsed far-ultraviolet (193 nm) laser radiation: Dependence of etch depth on experimental conditions. J Polym Sci Polym Chem Ed 1984; 22:2601-9. Taboada J, Mikesell GW Jr, Reed RD. Response of the comeal epithelium to KrF excimer laser pulses. Health Phys 1981; 40:67783. O'Brien WJ. Measurement of comeal DNA content. Invest Ophthalmol Vis Sci 1979; 18:538-43. Patrick MH. Physical and chemical properties of DNA. In: Wang SY, ed. Photochemistry and Photobiology of Nucleic Acids. New York: Academic Press, 1976; Vol 2:22. Loofbourow JR, Gould BS, Sizer IW. Studies of the ultraviolet absorption spectra of collagen. Arch Biochem 1949; 22:406-11. Maurice DM. The comea and sclera. In: Davson H, ed. The Eye, 3d ed. New York: Academic Press, 1984; Vol. 1b:12. Stone AL. Optical rotary dispersion of mucopolysaccharides. III. Ultraviolet circular dichroism and conformational specificity in amide groups. Biopolymers 1971; 10:739-51. Stone AL. Personal communication. Epstein WL, Fukuyama K, Epstein JH. Early effects of ultraviolet light on DNA synthesis in human skin in vivo. Arch Dermatol 1969; 100:84-9.